Essentials

To understand the rationale for the importance of extracellular vesicles (EVs) and its importance in early-stage cancer detection, it is essential first to understand the four main types of biomolecules and generally understand how cancer forms. There are four main biomolecule types: carbohydrates, fats/lipids, proteins, and nucleic acids. Carbohydrates are primarily used for generating and storing energy in the human body, cell adhesion, and cell-cell recognition. Like carbohydrates, fats are essential for energy storage, cell structure, insulation, organ protection, and steroid hormones. Proteins are essential for creating enzymes, building and repairing tissues, transporting important molecules throughout the human body, hormones, and immune function. Nucleic acids are the primary information holders within a cell. As famously stated by Watson and Crick, the central dogma of Molecular Biology is that DNA is transcribed to mRNA, and mRNA is translated to protein. The core information holder within an animal cell is DNA.

Cancer, at its core, is caused by mutations in a cell's DNA that lead to excessive cell growth and duplication, an inhibition of genes that control apoptosis, and/or a loss of function mutation on DNA caretaking genes. Any unmutated gene that promotes cell growth is referred to as a proto-oncogene. These gene types normally are expressed or inhibited in a controlled manner by the cell. However, some mutations can cause continuous gene expression (which would now be categorized as an oncogene) and lead to uncontrolled cell proliferation. The cell's caretaking genes must function normally to repair or correct these damaged sections of DNA. Otherwise, a loss of function mutation in these genome repair genes can increase the odds that the cell continuously proliferates. If tumor suppressor genes—such as apoptosis-inducing genes—do not kill the cell, it is up to the immune system to potentially stave off cancer proliferation. Unfortunately, the immune system is not always effective.

Why RNA and DNA-Centric Assays Have Had Subpar Performance

Considering cancer is the result of mutated DNA, I am not surprised that billions of dollars have been invested into making cancer diagnostic tests that are heavily based on information given from DNA and RNA. Companies focused on DNA-based methods only analyze point mutations, DNA length, and methylation patterns within circulating-free DNA (cfDNA)—small fragments of DNA floating around in biofluids that are composed of approximately 160 base pairs out of the 3 billion base pairs that comprise the human genome. Detecting DNA variation in early stages is too difficult with current sequencing bandwidth, a lack of cancer genome sequences identified for early-stage cancer, and a lack of cancer DNA in early and benign cancers. In most cases, less than 0.01% of the total cfDNA are circulating-tumor DNA (ctDNA) in early-stage cancer. The sparse and hard-to-identify ctDNA is especially problematic because methods used to analyze those fragments of DNA have an error rate well above 0.01%, and that error rate often leads to an abundance of false positives due to incorrectly analyzed DNA fragments.

RNA effectively has all the same limitations as DNA for early-stage cancer detection plus isolation-related degradation issues. RNA is single-stranded, so it is not as stabilized as DNA because DNA is double-stranded. DNA's second strand somewhat shields from conditions that can permanently change or break down the polymer. Single strand-based exposure is often a limitation that Isolation techniques have to overcome, but oftentimes damages the RNA and attenuates any cancer-predictive markers.

The Newly Emerging Paradigm is Still Off

The newly emerging dominant ideology for early cancer detection is that a multi-omic approach test will have the highest sensitivity and specificity for early-stage cancers. These people—who are reasonably competent—believe that analyzing more biomolecule types gives more information to analyze. Their rationale is that more information tends to correlate with improved test performance because more data sets have differential expression with cancer patients and healthy patients, for there are more thresholds or predictors to set for early-stage cancer detection within an algorithm. That cost-skyrocketing approach can seem to make sense if one does not consider the lack of conservedness and abundance of biomolecules related to cancer compared to EVs.

An Assay Centered on the Colocalization of Biomarkers on EVs is the Most Bayesianly and Biologically Sound Approach to Early Cancer Detection

It is crucial to efficiently analyze statistically sound biological molecule sources, so EVs must be the center focus for any test that wants to reach the apex of early-stage cancer detection. Let us apply Bayesian statistics to plasma. There are 10 billion EVs per mL of plasma. In contrast, about 500-1000 genome equivalents of cfDNA are in one milliliter of plasma. Using the knowledge that there are three nucleic acids per codon, the average protein is composed of 143 amino acids, and assuming that each codon codes for an amino acid, one can estimate that there are about 116.3 million protein equivalents (PE) worth of data in cfDNA per mL of plasma. In contrast, at minimum, there are 12 proteins per EV in plasma, so there are over 120 billion proteins within and in the surface of EVs per mL of plasma. Therefore, cancer-related markers are approximately 1000 times more abundant in EVs than in cfDNA.

The lack of colocalization-ability associated with circulating nucleic acids (cfNA) further caps the performance of cfNA-based tests relative to EV-based tests. cfDNA is roughly 167 base pairs long on average. That equates to about 56 PE units of data. In contrast, there are—at a minimum—1716 PE units of data in one EV if one merely looks at proteins on and within the EV's membrane. There is 30 times more data in just proteins within an EV. The abundance of information per EV allows well-designed and EV-based assays and their models to predict the cell of origin cell type from which an EV was generated. Easily obtainable origin information should theoretically boost the sensitivity and specificity of EV-based tests. On a biological level, EVs' cargo is well-protected and effectively unique to the cell of origin. These properties further improve the specificity of an EV-centric assay.